lex_node

Overview

The ROS lex_node node enables a robot to comprehend natural language commands by voice or textual input and respond through a set of actions, which an AWS Lex Bot maps to ROS messages. Out of the box this node provides a ROS interface to communicate with a specified Amazon Lex bot (configured via lex_config.yaml) and requires configuration of AWS credentials. The Amazon Lex bot needs to be defined with responses and slots for customer prompts. A set of default slots and mappings are demonstrated in the sample app and include actions as “Create ,” “Go to ” and “Stop.” Additional guides on configuring bots with are available at Getting Started with Amazon Lex.

Delivering a voice-enabled customer experience (e.g. “Robot, go to x”) will require dialog facilitation, wake word, and offline processing which are not yet provided by this integration. A wake word would trigger the dialog facilitation node to start recording and send the audio to Amazon Lex, then prompt the user for more information should Amazon Lex require it.

The ROS lex_node wraps the aws-sdk-c++ in a ROS service API.

Amazon Lex Summary: Amazon Lex is a service for building conversational interfaces into any application using voice and text. Amazon Lex provides the advanced deep learning functionality of automatic speech recognition (ASR) for converting speech to text, and natural language understanding (NLU) to recognize the intent of the text, to enable you to build applications with highly engaging user experiences and lifelike conversational interactions. With Amazon Lex, the same deep learning technologies that power Amazon Alexa are now available to any developer, enabling you to quickly and easily build sophisticated, natural language, conversational bots (“chatbots”).

License

The source code is released under an Apache 2.0.

Author: AWS RoboMaker
Affiliation: Amazon Web Services (AWS)
Maintainer: AWS RoboMaker, ros-contributions@amazon.com

Supported ROS Distributions

• Kinetic
• Melodic

Build status

• Travis CI:
  • ROS build farm:
  • ROS Kinetic @ u16.04 Xenial
  • ROS Melodic @ u18.04 Bionic

Installation

AWS Credentials

You will need to create an AWS Account and configure the credentials to be able to communicate with AWS services. You may find AWS Configuration and Credential Files helpful.

This node requires an IAM User with the following permission policy: - AmazonLexRunBotsOnly

Binaries

On Ubuntu you can install the latest version of this package using the following command

    sudo apt-get update
    sudo apt-get install -y ros-$ROS_DISTRO-lex-node

Building from Source

To build from source you'll need to create a new workspace, clone and checkout the latest release branch of this repository, install all the dependencies, and compile. If you need the latest development features you can clone from the master branch instead of the latest release branch. While we guarantee the release branches are stable, the master should be considered to have an unstable build due to ongoing development.

• Create a ROS workspace and a source directory

  mkdir -p ~/ros-workspace/src
  

• Clone the package into the source directory .

  Note: Replace _{MAJOR.VERSION}__ below with the latest major version number to get the latest release branch._

  cd ~/ros-workspace/src
  git clone https://github.com/aws-robotics/lex-ros1.git -b release-v{MAJOR.VERSION}
  

• Install dependencies

  cd ~/ros-workspace 
  sudo apt-get update && rosdep update
  rosdep install --from-paths src --ignore-src -r -y
  

Note: If building the master branch instead of a release branch you may need to also checkout and build the master branches of the packages this package depends on.

• Build the packages

  cd ~/ros-workspace && colcon build
  

• Configure ROS library Path

  source ~/ros-workspace/install/setup.bash
  

• Build and run the unit tests

  colcon build --packages-select lex_node --cmake-target tests
  colcon test --packages-select lex_node && colcon test-result --all
  

Launch Files

An example launch file called sample_application.launch is provided.

Usage

Resource Setup

1. Go to Amazon Lex
2. Create sample bot: BookTrip
3. Select publish, create a new alias
4. Modify the configuration file in config/sample_configuration.yaml to reflect the new alias

Run the node

• With launch file using parameters in .yaml format (example provided)
  • ROS: roslaunch lex_node sample_application.launch

Send a test voice message

`rosservice call /lex_node/lex_conversation "{content_type: 'text/plain; charset=utf-8', accept_type: 'text/plain; charset=utf-8', text_request: 'make a reservation', audio_request: {data: ''}}"`

Verify the test voice was received

• Receive response from Amazon Lex and continue conversation

Configuration File and Parameters

An example configuration file called sample_configuration.yaml is provided.

Client Configuration
Namespace:

Amazon Lex Configuration
Namespace:

Performance and Benchmark Results

We evaluated the performance of this node by runnning the followning scenario on a Raspberry Pi 3 Model B:

• Launch a baseline graph containing the talker and listener nodes from the roscpp_tutorials package, plus two additional nodes that collect CPU and memory usage statistics. Allow the nodes to run for 60 seconds.
• Launch the ROS lex_node using the launch file lex_node.launch as described above. At the same time, make calls to the /lex_node/lex_conversation service by running the following script in the background:

rosservice call /lex_node/set_logger_level "{logger: 'ros.lex_node', level: 'debug'}" 
rosservice call /lex_node/lex_conversation "{content_type: 'text/plain; charset=utf-8', accept_type: 'text/plain; charset=utf-8', text_request: 'Make a reservation', audio_request: {data: ''}}" && sleep 1
rosservice call /lex_node/lex_conversation "{content_type: 'text/plain; charset=utf-8', accept_type: 'text/plain; charset=utf-8', text_request: 'Seattle, WA', audio_request: {data: ''}}" && sleep 1
rosservice call /lex_node/lex_conversation "{content_type: 'text/plain; charset=utf-8', accept_type: 'text/plain; charset=utf-8', text_request: 'Tomorrow', audio_request: {data: ''}}" && sleep 1
rosservice call /lex_node/lex_conversation "{content_type: 'text/plain; charset=utf-8', accept_type: 'text/plain; charset=utf-8', text_request: 'Next Monday', audio_request: {data: ''}}" && sleep 1
rosservice call /lex_node/lex_conversation "{content_type: 'text/plain; charset=utf-8', accept_type: 'text/plain; charset=utf-8', text_request: '40', audio_request: {data: ''}}" && sleep 1
rosservice call /lex_node/lex_conversation "{content_type: 'text/plain; charset=utf-8', accept_type: 'text/plain; charset=utf-8', text_request: 'economy', audio_request: {data: ''}}" && sleep 1 
rosservice call /lex_node/lex_conversation "{content_type: 'text/plain; charset=utf-8', accept_type: 'text/plain; charset=utf-8', text_request: 'yes', audio_request: {data: ''}}"

• Allow the nodes to run for 180 seconds.
• Terminate the ROS lex_node, and allow the remaining nodes to run for 60 seconds.

The following graph shows the CPU usage during that scenario. The 1 minute average CPU usage starts at 15% during the launch of the baseline graph, and stabilizes around 9%. When we launch the lex_node node around second 85 the 1 minute average CPU increases up to a peak of 27.75%, and stabilizes around 24% while the lex_node node serves the service calls. After that the 1 minute average CPU usage is kept around 12% until we stop the lex_node node around second 360.

The following graph shows the memory usage during that scenario. We start with a memory usage of around 253 MB for the baseline graph, that increases to a around 280 MB (+10.67%) after launching the lex_node node around second 85, and increases again to a peak of 300 (+18.58% wrt initial value) while the lex_node node serves the service calls. After that the memory usage decreases to around 278 MB, and keeps this way until going back to 253 MB after we stop the lex_node node.

Node

lex_node

Enables a robot to comprehend natural language commands by voice or textual input and respond through a set of actions.

Services

Topic: ~/lex_conversation

AudioTextConversation

Request:

Response:

Subscribed Topics

None

Published Topics

None

Bugs & Feature Requests

Please contact the team directly if you would like to request a feature.

Please report bugs in Issue Tracker.